Background: Endoscopic optical coherence tomography (EOCT) is a high-resolution, cross-sectional tissue-imaging technique that provides microscopic morphologic information. EOCT should detect dysplasia in Barrett's epithelium, but this has not been established in a prospective blinded study. This study evaluated the accuracy of EOCT for the diagnosis and the exclusion of dysplasia in patients with Barrett's esophagus.
Methods: A 2.4-mm diameter EOCT probe was modified for use with a cap-fitted, two-channel endoscope. Pairs of EOCT image streams and jumbo biopsy specimens were obtained. Endoscopy/EOCT procedures were performed by 4 endoscopists who separately reviewed the EOCT digital images for the absence or the presence of dysplasia (low grade, high grade, or cancer) for each biopsy specimen obtained. The endoscopists were blinded to the interpretation of the pathology. An experienced pathologist blinded to the endoscopic/EOCT findings evaluated each biopsy for the absence or the presence of dysplasia. The setting of the study was a major academic medical center. Adult patients with documented Barrett's esophagus greater than 2 cm were included in the study. The main outcome measurement was the accuracy of EOCT in the detection of dysplasia in patients with Barrett's esophagus.
Results: A total of 314 usable EOCT image stream/biopsy pairs were obtained in 33 patients. By using histology as the standard, the performance of EOCT was sensitivity, 68%; specificity, 82%; positive predictive value, 53%; negative predictive value, 89%; and diagnostic accuracy, 78%. Diagnostic accuracy for the 4 endoscopists ranged from 56% to 98%. Limitations of the study were the variability in endoscopists' accuracy rates, difficulty in real-time interpretation, and the need for refined criteria of dysplasia by EOCT imaging.
Conclusions: The current EOCT system has an accuracy of 78% for the detection of dysplasia in patients with Barrett's esophagus. EOCT could be used to target biopsies to areas of Barrett's epithelium with a higher probability for the presence of dysplasia. However, further modifications, including increased resolution and identification of further potential OCT characteristics of dysplasia, are needed before EOCT can be used clinically.